The orthodoxy this plate inverts was set down in the volume's second plate, and it is worth restating in its full serenity, because the force of the inversion is measured against it. In a labour market of many buyers and many sellers, no single one of them large enough to move the going wage, the employer takes that wage as given and hires until the value added by the last worker has fallen just to meet it. The wage equals the marginal product. The worker is paid, to the penny, the worth of the final hour. It is a clean picture, and a flattering one, and for the better part of a century it was taught much as one teaches the weather — as a thing that simply obtains, prior to anyone's will.
Monopsony is its mirror image. Suppose there is not a crowd of employers but one; or a few; or — as we shall come to see is the decisive case — merely a market sticky enough that a worker cannot walk away from a poor wage without paying for the privilege. Now the lone buyer confronts the whole upward-sloping supply curve of labour. To draw in one more worker it must raise the wage, and it must pay that higher wage not to the newcomer alone but to everyone already on the books. The cost of the marginal worker therefore stands above that worker's wage. And the employer, hiring as any optimizer does up to the point where the marginal cost of labour meets labour's marginal product, stops short — engaging fewer hands than a competitive market would, and paying each of them less than the value of what they make.
The space between the wage and the worth is not friction, and it is not an accident of measurement. It is the measure of the employer's power.
The space between the wage and the worth is not friction, and it is not an accident. It is the measure of the employer's power.
Joan Robinson gave the thing its name in 1933, in the same book that gave imperfect competition its own, and she had the word from a Cambridge classicist who assembled it, a little absurdly, from the Greek for a single buyer of fish. The absurdity is fitting, for the concept is monopoly's neglected twin, and twins are often mistaken for curiosities. Where the lone seller holds the price up, the lone buyer holds the wage down; the geometry is the same diagram held to a glass. The fuller anatomy of that power — its forms, its measures, its mirror in the market for goods — belongs to a later volume, where monopsony becomes the whole body. Here it is one organ only: the buyer of labour, and what its power does to the wage.
For a long time this was filed under curiosities. The textbook's monopsonist was the isolated company town — the one mine, the one mill, the worker with nowhere else to carry his labour and no fare to get there if there had been somewhere to go. A real case, and a vivid one, and on inspection a rare one: a corner of the world, a footnote to the competitive rule. Most workers, after all, could choose among employers; most labour markets held many buyers; and so the orthodox picture stood, and the wage measured worth, and the company town remained the exception that flattered the rule by being so plainly exceptional.
From this followed a confident corollary about the minimum wage, set out in the plate just before this one. If the wage already equals the marginal product, then a legal floor raised above it must price the least productive workers out of their posts: the employer, forbidden to pay them what they are worth, will simply let them go. The minimum wage was, on this reading, a tax on the employment of the poor, and a peculiarly cruel one, since it fell hardest on those whose labour commanded least. This was no fringe view held by cranks. It was the settled teaching of the profession, repeated in every introductory lecture hall, and it carried the considerable authority of following rigorously from a model that nearly everyone believed.
And yet the same model, looked at squarely, contains a result that ought to have unsettled the confidence long before it did. Because the monopsonist is already holding employment below the level competition would set, a minimum wage — pitched above the wage it pays, but below the competitive wage — can take from it the very instrument of its restraint. Told that it may pay no less than the floor however few it hires, the employer no longer saves anything by hiring few. The marginal cost of an additional worker collapses to the floor itself, flat; the wedge that made restraint profitable is gone; and the firm expands its hiring toward the level competition would have reached. The wage rises, and employment rises with it. The very policy the textbook held must destroy jobs can, where the buyer has power, be the thing that makes them.
This is no heterodox sleight of hand. It is ordinary price theory, plainly implied by Robinson's own diagram and worked through, with some care, by George Stigler — no friend of minimum wages, and no one's idea of a radical — as early as 1946. For decades the result lived a curious double life: conceded in the seminar room, where its logic was unanswerable, and waved away in practice, where everyone simply knew that real labour markets were competitive and so the monopsony case did not arise. The theory was admitted. The world was assumed not to contain it.
The theory was admitted. The world was assumed not to contain it.
Then the world was asked. In 1992 New Jersey raised its minimum wage and Pennsylvania, across the river, did not; and David Card and Alan Krueger went to the fast-food counters of both states to count who was working. The competitive model made a clean and falsifiable prediction: employment in New Jersey should fall, relative to its neighbour. It did not. By their measure, if anything, it rose. The finding, published in 1994, fell on the profession like a dropped tray in a quiet dining room.
The reaction was ferocious, as reactions to overturned orthodoxies tend to be. The data were contested — payroll records set against telephone surveys, with David Neumark and William Wascher pressing hard for the other numbers; the methods were turned over; the result was qualified, replicated, and extended across hundreds of later studies in many countries. The edges of that literature remain genuinely unsettled, and an honest plate says so plainly rather than claiming a victory the evidence has not fully delivered. But the centre has largely held: modest increases in the minimum wage, raised from low levels, do not produce the sharp loss of jobs the competitive model foretold. Card would share the Nobel in 2021, his half cited for the empirical study of labour — for, among much else, teaching the discipline to let the world correct the diagram. And the explanation that fits what was found most naturally is precisely the one the textbook had kept folded in a drawer: where employers hold power over the wage, the competitive prediction does not apply.
And still the company town is rare. Had monopsony required a single literal buyer, it would have stayed the footnote it was always taken to be. The decisive move — the one that turned monopsony from curiosity into a general account of the labour market — was the recognition that one needs no single buyer at all. One needs only friction.
Consider what it truly costs a worker to leave a low wage. The search for another post, conducted in the evenings, against incomplete knowledge of what is even out there. The risk of the unfamiliar employer who may be worse. The seniority forfeited, the pension vesting reset, the family uprooted, the commute lengthened, the particular shift that happened to fit the school run and will not be found again easily. Each such friction, and there are always some, hands the employer a sliver of power over the wage — because a worker who cannot leave without cost can be paid somewhat under his worth and still, on balance, stay. The market need not be a company town. It need only be sticky. Alan Manning gave this its modern statement in 2003, under the name of dynamic monopsony, and in doing so changed the question. No longer "is there a single employer?" but "how readily does labour flow toward a firm that lifts its wage, and away from one that cuts it?" — the elasticity of labour supply to the individual firm. Where that elasticity is high, the firm is a price-taker and the old orthodoxy holds. Where it is low, the firm has power. And the accumulating evidence — drawn from how quickly workers quit when their pay falls behind, from how much harder a firm must work to recruit when it offers less — is that it is, very often, low. The labour market is not a competitive expanse dotted with the odd company town. It is monopsonistic, in a thousand small degrees, very nearly everywhere.
Once the eye is trained to it, a whole architecture of wage-suppression swims into view — and each beam of it is, at bottom, a device for lowering the elasticity of labour supply to the firm. Where a few employers dominate a local market — a couple of hospital systems across a county, a single vast warehouse on the edge of a town — wages run lower, and the measured concentration of buyers tracks the suppression with some fidelity. The non-compete clause, once the lawyer's instrument for restraining a departing executive, now binds the sandwich-maker and the warehouse hand, forbidding the worker the one move that would have disciplined the wage. The no-poach agreement — the technology firms who quietly undertook not to hire one another's engineers; the franchise contracts that barred one outlet from engaging another's staff — is collusion on the buying side, monopsony assembled by treaty. Occupational licensing raises the toll for carrying a skill across a state line. Even the plain opacity of pay, the lingering taboo on saying aloud what one earns, is a friction that works for the buyer. None of these is a blemish on an otherwise competitive market. Each is a way of manufacturing the stickiness that gives the wage its slack — of widening, deliberately, the gap between what the worker is paid and what the worker is worth.
Here the honest plate must mark the limits of its own case — for monopsony has lately become an orthodoxy in its turn, and an orthodoxy congenial to one's politics is the most dangerous kind there is. How much power, on average, do employers hold? The honest answer is that we are still measuring, and that the firm-level elasticity of labour supply is hard to pin down and surely varies enormously: the nurse in a rural county is held far more tightly than the programmer in a great city, and to speak of the degree of monopsony is already to smooth flat the thing that matters most. The minimum-wage evidence carries a range it does not entitle us to leave behind: a floor raised gently from a low base is a different object from a floor set near half the median wage, and the cushion monopsony provides is finite — press the wage high enough and even a buyer with power will, in the end, hire fewer. The New Jersey result was a result about a particular rise, in a particular trade, from a particular level. It is not a warrant for the proposition that wages may be legislated upward without cost, and the careful reader should distrust anyone who wields it so.
What monopsony establishes is at once narrower and more durable than the slogans pinned to it from either side. It establishes that the competitive model's confident prediction of lost jobs does not hold where buyers have power; and that buyers, far more often than the old teaching allowed, do. It does not establish that the labour market is everywhere a company town — any more than the orthodoxy it unseated had established that it was nowhere one. The truth, very nearly certainly, is that the market is competitive in some places and monopsonistic in others, in proportions and by degrees that the discipline is still, slowly and ungloriously, learning how to read. That admission is not a weakness of the account. It is the account, told honestly.
Return, then, to the question this volume will not stop asking, and that the whole table was built to put to every price: is the wage the measure of the worker's worth, or of the employer's power? The marginal-productivity orthodoxy gave the clean reply — worth, and worth alone — and raised upon it a vision of the labour market as a place where each is paid exactly what he adds, no more and no less, as if by a law of nature. Monopsony returns the harder reply: in part, power; and the part is not small, and it is not confined to the mining town. The wage is the settlement of a contest — between the ease with which a worker might leave and the many small contrivances that make leaving dear, between the outside option and the grip upon it.
To raise a minimum wage, then; to void a non-compete; to bring the antitrust laws at long last to the buying side of the labour market — this is not, or not merely, to meddle with a price that nature quietly set. It is to alter the terms of that contest, to loosen the grip by a degree; and the smoothness of the quoted wage was always, in part, a way of concealing that there was a contest at all. That is the inversion. It is why this plate stands eighth of the volume's twelve, at the hinge on which The Wage turns from the worker's worth toward the employer's power — and why the question it leaves open is the very one the table itself closes upon: not whether the market is free, but in whose favour it was built.